Treatment of atrial fibrillation by cryoablation of the pulmonary vein (PV) suffers from an inability to assess probe contact, tissue thickness, and freeze completion through the wall. Unfortunately, clinical imaging cannot be used for this purpose as these techniques have resolutions similar in scale (~ 1 - 2 mm) to PV thickness and therefore are unable to resolve changes within the PV during treatment. Here, a micro-thermal sensor based on the "3?" technique which has been used for thin biological systems is proposed as a potential solution and tested for a cryoablation scenario. First, the sensor was modified from a linear format to a serpentine format for integration onto a flexible balloon. Next, using numerical analyses, the ability of the modified sensor on a flat substrate was studied to differentiate measurements in limiting cases of ice, water and fat. These numerical results were then complemented by experimentation by micro-patterning the serpentine sensor onto a flat substrate and onto a flexible balloon. In both formats (flat and balloon) the serpentine sensor was experimentally shown to: (1) identify tissue contact vs. fluid, (2) distinguish tissue thickness in the 0.5 to 2 mm range, and (3) measure the initiation and completion of freezing as previously reported for a linear sensor. This study demonstrates proof of principle that a serpentine 3? sensor on a balloon can monitor tissue contact, thickness and phase change which is relevant for cryo and other focal thermal treatments of PV to treat atrial fibrillation.